Applications of ACD/NMR Software Suite Applications for the identification and databasing of chemical markers in biological fluids

by Chris Silwood, Research Fellow, Inflammation Research Group, St. Barthlomew's and the Royal London Hospitals School of Medicine and Dentistry, London E1 2AD.

Along with our research director, Martin Grootveld, I employ NMR (1H, 13C, 19F, 31P) spectroscopy as a means of identifying chemical markers in a wide variety of biological fluids. These can be the well-resolved, sharp signals attributable to low molecular mass metabolites as well as mobile protons contained within macromolecules. In particular, projects consider the deleterious effects of free radical reactions in vivo, for conditions such as rheumatoid arthritis. 1H NMR is especially useful for the detection of free radical-induced lipid oxidation products (LOPs), although there can be a great deal of diagnostic unsaturated proton signal overlap, and even apparent chemical equivalence of spins. This requires the use of 2-dimensional techniques, hence we employ, on a regular basis, 1H-1H COSY, RCT, TOCSY, JR and 1H-13C HMQC/HSQC methods for the detection of such products.

Trial assignments and extracted spectral parameters require confirmation and we have found the ACD/HNMR and ACD/CNMR products to be invaluable for this purpose. Predictions of, in particular, chemical shifts, provide the basis for confirmation of tentative assignments based on a consideration of substituent effects, etc. The extracted experimental spectral parameters can thus be used to fine-tune the predicted ACD shifts, the appearance of the newly-generated multiplets after a new prediction calculation serving as a final confirmation of correct assignment. The employment of this strategy is ideally suited to the software suite because of the mixture facilities, as we rarely detect less than four oxidized moieties in the unsaturated proton region. The application of this methodology to the study of the LOPs resulting from the heating of culinary oils (cf. Grootveld et al., J.Clin.Invest. 101 (1998), 1210-1218, http://www.jci.org/cgi/content/full/101/6/1210) is expected to be published shortly, in conjunction with ACD Inc.

Concerning our research in general, as well as identifying particular markers and metabolites within fluids, new, unidentified signals in spectra have to be identified. This is a necessary consequence of the appraisal of patient-specific data and can lead to spectra containing information for some fifty compounds with the inherent complexity of signal overlap that that brings. For this reason, a rigorous method for generating predicted spectra has to be fast, the CPU time not being limited by size of molecule (e.g. C18 systems within the lipid class) due to the number of suspected derivatives that have to be screened for spectral properties. To this end, we find the database module of the ACD products equally invaluable for building up compendia of our spectra that specifically contain information on macroscopic aqueous medium shielding effects and any further effects resulting from component interactions within these complex mixtures, as well as any further macroscopic effects reflecting electrolyte levels. The larger the database within, for example, the HNMR suite, the greater the influence on subsequent shift calculations of new compounds. We are also applying the database strategy to experimental spectra, which brings in the unique capability (in its simplicity) of the ACD software for searches, for specific compound types, or just typical fragments therein.

Based on fruitful correspondence with ACD Inc., we intend to make available in the near future our experimental fluid databases for incorporation into ACD/NMR Manager for interested parties and back this up with a further pH database for appropriate molecules that exhibit sensitivity to pH, as a further indicator of the intermolecular interactions that restrict a fluid to a particular pH range.